One of the core aspirations of Industry 4.0 is to have machines that predict when they will fail. The goal is that small sensors embedded within components such as hydraulic systems or fans will determine when an issue will occur. By discovering faults before they happen, minor issues can be resolved before they cause major problems. For example, in a airplane, if a component fails outside of regular maintenance schedules, it will cause the plane to be taken out of service therefore resulting in delays and lost of cost issues. Although some forms of predictive maintenance sensors, such as vibration monitoring systems, have been around for decades, others have been slower to evolve.
One of those starting to evolve is the prediction of the failure of fans. Fans are in many things from computers to HVAC systems and their failure can cause all sorts of problems. By analyzing the sound that a fan is making Augury can predict when the fan will fail.
This small device records the vibrations and ultrasonic sounds of a fan and then uploads the information to Augury's cloud service which analyzes it and then makes predictions about the state of the machine. Once uploaded, engineers can view the status of the machine and see any alerts using the company's app.
There is of course a problem, how does Augury know what the sound means? That's the idea of storing the sounds on a cloud service. Initially, once the problem was found, it's hoped that technicians will label their sound on the cloud so that, after several similar sounds, it will recognize all future sounds and alert new users who upload their noises. By crowd-sourcing these sounds, the data-based will increase to such a point that where any noise that a fan makes can be diagnosed immediately.
It's still early for Internet of Things (IoT) based predictive maintenance, but Augury shows the potential for such devices and the need for crowd-sourced services in which engineers in similar industries can access information about devices they maintain and fix problems before they occur.